December 1999 - Zope-Dev

Eliminating forced HEAD/TITLE/BASE HREF
by Tom Neff 07 Dec '99

07 Dec '99

I need to use Zope output in a server-side include on another machine that's not running Zope. Has anybody come up with a reasonably clean way to suppress the enforced emission of the opening HTML, HEAD, TITLE and BASE HREF tags? It is rather frustrating, given the package's great sophistication otherwise, that one doesn't have control over those odd bits of HTML. I looked at the Python code but it seems to be the usual multilayered marvel and I can't quite see where to put the scalpel. Any assistance appreciated.

2 2

mod_pcgi2 version 0.0.1
by Oleg Broytmann 07 Dec '99

07 Dec '99

Hello! I am announcing the first development version of mod_pcgi2 - Apache module for Zope/PCGI. The module is at version 0.0.1 - it is very early development release. It lacks some features and probably is buggy (though all my tests passed fine). http://www.zope.org/Members/phd/mod_pcgi2 http://sun.med.ru/~phd/Software/Zope/ I've tested it on Debian Linux and Sun Sparc Solaris. Here are my pages: http://sun.med.ru/pcgi/ http://sun.med.ru/pcgi/Random/ If you suspect there is a problem, try to view these pages by replacing /pcgi/ with /Zope/ - this will force Apache to use Zope.cgi. My home page (http://sun.med.ru/~phd/) loaded into Zope: via Zope.cgi: http://sun.med.ru/~zphd/ via mod_pcgi2: http://sun.med.ru/pcgi/phd02 Zope + PostgreSQL little test: via Zope.cgi: http://sun.med.ru/Zope/phd01 via mod_pcgi2: http://sun.med.ru/pcgi/phd01 On linux: via Zope.cgi: http://phd.russ.ru/Zope/ via mod_pcgi2: http://phd.russ.ru/pcgi/ Please visit (some of) these pages and report any error or success. I'd be very happy if any Apache guru comments on the mod_pcgi2 code. Oleg. ---- Oleg Broytmann Foundation for Effective Policies phd(a)phd.russ.ru Programmers don't die, they just GOSUB without RETURN.

1 0

Understanding Roles?
by Jim Penny 07 Dec '99

07 Dec '99

Hi: This messages has several attached html files. I have defined a folder; GAP, and a local role; UserAdministrator. In GAP/acl_users, I have 2 users, julian, who has the UserAdministrator role, and betty who does not. See attachment users.html In folder GAP, UserAdministrator has the roles given in users0.html. To summarize, all Add_.. and Change_.. have been removed from Acquire. UserAdministrators are given roles "Access Content Info", "Change Configuration", "Change Proxy", "Change Permissions", "Manage properties", "Manage users", "Use Database MEthods", "Use mailhost services", "View", and "View management screens" Some of these rights in fact make me nervous, and I would like to trim more. See attachment "users0.html". To confirm that julian has the rights of a UserAdministrator, see attachment "users1.html". Now, in folder GAP/acl_users, UserAdministrators are granted all roles. except manage properties and Undo changes. See attachment "users2.html". In particular, UserAdministrators are given the role of "Access Content Info". But, when I log in as "julian", I and go to GAP/manage, and click on acl_users, I am presented with a menu which _DOES NOT_ show the contents (users) of the folder; instead I am presented with a tabbed menu that has only "Security" in it. See attachment "users3.html". Why does "julian" not have a "Contents" tab in GAP/acl_users? Any idea of what the minimal set of permissions will allow me to create a local role that permits members to add, set passwords, and delete users from a pre-existing acl_users folder?

1 0

Reason for Uathorization Failure
by Jim Penny 06 Dec '99

06 Dec '99

How does one determine the reason that authorization is denied to a particular resource? Is the fact of denial logged anywhere? I am trying to set up two local roles for some projected sections of a website. I would like to create a UserAdministrator local role who can add, delete, or modify user passwords, and a Browser role, who must log in, but may not do anything but view. I am having trouble getting the permissions right for this scenario. Any hints? Thanks. Jim Penny

1 0

"Official" solution for Squishdot and 2.1.0 ?
by Gregor Hoffleit 06 Dec '99

06 Dec '99

Last Thursday, Scott Kaplan posted a hack to SearchIndex/TextIndex.py to make Squishdot 0.3.2 work with Zope 2.1.0, asking for a proper fix to this problem. The problem looked quite trivial and certainly Squishdot.py should be fixed instead, but I fail to see how and wonder if some expert is looking into this. Gregor

1 0

Zope guru help needed (or become guru if you help me)
by Martijn Faassen 05 Dec '99

05 Dec '99

Hi there, Yesterday I released ZFormulator 0.1. I've run into a problem with Zope security settings with it now. I would appreciate it a lot if anyone knowledgable and daring enough would take a look at the file Form.py in ZFormulator. The symptoms are as following: I've been developing ZFormulator logged in as the Zope superuser. Apparently this Zope superuser has powers above the manager role, as ZFormulator works for the superuser, but not for the manager. I don't know exactly why, but I do know I am doing a lot of hacking in Form.py that might be causing it. Here I'll describe some of the hackery. It is probably completely incomprehensible if you haven't at least tried to read Form.py in ZFormulator first (or are reading it while you read this post). Also check out __init__.py to see how Form.py is used from the other side. In order to be able to easily register new field classes with Zope, without having to do a lot of boring code duplication (which would mean also maintenance hassles), I've tried to shield the Python programmer as much as possible from Zope itself. Form.py contains, besides the code implementing ZFormulatorForm, also code to register fields with it. The normal way to register a product in Zope is to tell Zope about a manage_add method, and a manage_addForm. In ZFormulator I only have a single manage_add method (which receives from the manage_add Form which field to add), and I dynamically add manage_addForm methods for each registered field (in the form manage_add<field meta_type>Form). These Forms cannot be HTMLFile objects, because they should be rendered differently for each field that is added. Because I could not figure out a way to pass extra information to the HTMLFile object when it's called, I decided not to use them, and wrote a custom class instead to produce the right form. Unfortunately, a mechanism in OFS/CopySupport.py needed when doing a copy & paste or a rename of an object in the Zope management interface, didn't like my custom class. It was looking for some mysterious (to me) attribute __roles__, and didn't find it. So while I was doing hacks anyway, I manually added in a __roles__ attribute to the FormHack class. After that renaming and copy and pasting *appeared* to work, though I've had some odd results after a rename recently, so I'm not entirely sure. This is one possible cause of the bug. A less hackish solution to all this would be appreciated, in any case. There's another possibility, however. Because the manage_add<meta_type>Form methods are added dynamically to the Form class, I can't include these Forms in the __ac_permissions__ of that class. So, I reasoned perhaps the superuser isn't bothered by this lack of permission information, but normal managers are, and thus can't access the form because they have no permission to do so. It seems like this is not the case however; I added yet more hackery stuff to the registration code, this time modifying the __ac_permissions__ attribute. If my hack went right the Forms *should* be part of __ac_permissions__ now and the manager ought to have no problems. But the problem persisted, so I don't think this is the solution. In general, I'd like to know if there's some better way to accomplish what I want without resorting to such hacking. To make my framework work well currently I *do* need it; I don't want to have to write an identical add<meta_type>Form.dtml file for each field that I add, I don't want to have to edit the __ac_permission__ attribute of the Form class for each field I add, I don't want to have to add manage_addForm methods to the Form class manually. Right now the framework works pretty nicely, shielding you from all that. Unfortunately it doesn't work properly yet... So, Zope gurus, *please* *please* take a look at Form.py. Please save me from all these kludges! Regards, Martijn

2 2

Rename non-funtional
by Bill Anderson 04 Dec '99

04 Dec '99

Uhhhhmmmm ... after repeated trys (closing browser, sevrer restart, etc) I am unable to rename _anything_. It returns the page telling me to select something first, and yes, I did select stuff. 2.1.0 is the only change I've made ... Bill -- In flying I have learned that carelessness and overconfidence are usually far more dangerous than deliberately accepted risks. -- Wilbur Wright in a letter to his father, September 1900

2 1

Re: [Zope-dev] how to install a method hook
by Alexandre Ratti 04 Dec '99

04 Dec '99

Hello Jephte, I cannot answer your question; however, if you're interested in internationalization issues (such as providing French date formats) you may want to take a look at the ZIP Project. The mailing list address is "zip(a)eevolute.com". To register, see the <http://www.eevolute.com/mailman/listinfo/zip> page. Cheers. Alexandre At 12:00 03/12/1999 -0800, you wrote: >Date: Fri, 03 Dec 1999 10:26:28 +0400 >From: Jepthte CLAIN <minf7(a)educ.univ-reunion.fr> >To: zope-dev(a)zope.org >Subject: [Zope-dev] how to install a method hook > > >Hello, > >here is the 'problem': I want to write support for french dates and >various other stuff in my zope installation. > >Until now, i've always patched zope, but it proves to be a pain to patch >my installation each time a release is out. > >for now, I want to write a product that 'replace' some common objects by >a my custom objects. > >For example, I derived DateTime to provide french date display and >input. I would like to 'replace' (in term of acquisition) the toplevel >Application.ZopeTime with my own ZopeTime (instance of DateTimeFr) <snip>

1 0

RE: [Zope-dev] Sybase DAv2 connection problems.
by Chris McDonough 04 Dec '99

04 Dec '99

I just tried connecting to a Sybase ASE 11.5 server (WinNT) with the SybaseDAv2 database adapter on Solaris. Before I tried to connect to it, a database was created named test (via Sybase Central "New->Database", default options) with no tables. I could connect to it without a problem. Weird. Chris McDonough mailto:chrism@digicool.com Digital Creations http://www.digicool.com Publishers of Zope http://www.zope.org > -----Original Message----- > From: Chris McDonough > Sent: Monday, November 29, 1999 5:11 PM > To: 'Bill Anderson'; Chris McDonough > Cc: zope-dev(a)zope.org > Subject: RE: [Zope-dev] Sybase DAv2 connection problems. > > > I will check this out.... thanks for the update! > > Chris McDonough > Digital Creations, Inc. > Zope - http://www.zope.org > > > > -----Original Message----- > > From: Bill Anderson [mailto:banderson@boi.hp.com] > > Sent: Monday, November 29, 1999 4:56 PM > > To: Chris McDonough > > Cc: zope-dev(a)zope.org > > Subject: Re: [Zope-dev] Sybase DAv2 connection problems. > > > > > > Just found it. > > > > Seems it doesn't like to connect to databases without tables ... go > > figure. > > > > -- > > Bill Anderson Linux/Unix > > Administrator, Security Analyst > > ESBU (ARC) bill_anderson(a)boi.hp.com > > My opinions are just that; _my_ opinions. > > > > _______________________________________________ > > Zope-Dev maillist - Zope-Dev(a)zope.org > > http://lists.zope.org/mailman/listinfo/zope-dev > > No cross posts or HTML encoding! > > (Related lists - > > http://lists.zope.org/mailman/listinfo/zope-announce > > http://lists.zope.org/mailman/listinfo/zope ) > > >

2 1

NOTICE about linux 2.1.0 binary, was: problems running 2.1.0 fin al al
by Brian Lloyd 03 Dec '99

03 Dec '99

> I have the same trouble trying to install the XMLDocument > product. My guess is > that somebody upgraded to a glibc2.1 system (e.g., RH 6.0) > and built the Linux > binaries (the .so files) there. We recently had to rebuild the binaries for the linux dist. (for all of them actually, to make sure that _localemodule was included by default). At the time, I didn't realize that the dev. machine had been "upgraded" to glibc2.1 :( The upshot is that the linux binary currently on the website is currently only usable for RH 6.0 users. If you _must_ upgrade before Monday and you are on 5.x, you will want to get the source distribution and build it yourself. Monday we will get the binaries for the linux binary dist. rebuilt on a 5.x machine and do some testing. Then I'll update the release on the site and we'll hopefully have a binary that's good for both 5.x and 6.x. Brian Lloyd brian(a)digicool.com Software Engineer 540.371.6909 Digital Creations http://www.digicool.com

1 0